OpenInference instrumentation for LangChain
An OpenTelemetry instrumentor for following LangChain application steps in a trace.
Published by Arize AI
How to use it
Use a small chain that retrieves a passage and calls a model. Check that the chain and its child spans share the same trace ID, then compare step duration and errors between two inputs.
Run pip install openinference-instrumentation-langchain in the environment running LangChain and configure a tracer provider. Check the installed integration’s compatibility before adding it to an existing application.
- Compatibility
- Python, LangChain, OpenTelemetry
- Monitoring areas
- Chain tracing, Retrieval spans
- Access
- Local Python instrumentation. A compatible OpenTelemetry receiver is needed to view traces.